Understanding Cooperation through Behavioural Economics
VOICEOVERWelcome to Melbourne University Up Close, a fortnightly podcast of research personalities and cultural offerings of the University of Melbourne, Australia. Up Close is available on the web at upclose.unimelb.edu.au. That’s upclose.unimelb.edu.au.
JENNIFER COOK
Hello, and welcome to Up Close coming to you from the University of Melbourne, Australia. I'm Jennifer Cook and in this episode of Up Close we will attempt to understand what drives humans to behave the way that they do.
To help us in our quest, we’re turning to what may seem the unlikely realm of economics. Here with me in the studio is Dr Nikos Nikiforakis, a behavioural economist and a lecturer at the University of Melbourne’s Department of Economics. Welcome to Up Close, Nikos.
NIKOS NIKIFORAKIS
Thank you Jen.
JENNIFER COOK
Let me first begin with a quote from economist and former chairman of the Federal Reserve of the United States, Alan Greenspan. In his testimony at the congressional hearing on October 23, 2008, he said this, “I made a mistake in presuming that the self interest of organisations, specifically banks and others, were such that they were best capable of protecting their own shareholders.”
Nikos, perhaps if you could take us now to your field which is behavioural economics and explain how this comment is such a huge shift from what went before.
NIKOS NIKIFORAKIS
Well it’s a very interesting question. So traditionally most economists, at least the last 100 years, have been working assuming that all individuals are fully rational and self interested, and behavioural economics is basically incorporating insights from psychology, social psychology, anthropology, biology and are trying to change these fundamental assumptions and by doing this, of course, we’re changing the way we view the world and what we believe is going to happen in the future.
JENNIFER COOK
So I’ve always thought of economic theory as quite separate from day to day human behaviour, so what has been the underlying rationale for economic theory up until now?
NIKOS NIKIFORAKIS
Economics is a social science so in that sense we’re interested in social issues - how people make decisions in day to day life. And of course our methodology differs from the other social sciences but at the end of the day we’re just trying to understand human decision making and how this affects social behaviour.
JENNIFER COOK
So, Nikos, how do we know how people behave? How do we measure that?
NIKOS NIKIFORAKIS
That’s a very good question. So if you think of Greenspan’s quote you would have thought such an educated person would have known. It’s only the last 60 years that economists have developed a methodology to conduct controlled laboratory experiments. Now as other social scientists we tended to restrict ourselves to just observation of people, create theories, trying to understand what we’re doing, try to predict behaviour but with the help of a strict mathematical theory in the last 60 years we have this methodology of how now to conduct lab experiments.
JENNIFER COOK
So you’re saying as if it was very helpful or not terribly helpful at all, quite restrictive?
NIKOS NIKIFORAKIS
It has been very helpful actually. A strict mathematical theory helps us generate accurate predictions and therefore we can actually test our theory whereas before we just couldn’t. There were so many possibilities. So now we have predictions we can see whether they’re accurate or inaccurate and the way we see this is by doing laboratory or sometimes field experiments and then we can go back to the theory and say, “well if it failed, if the prediction wasn’t accurate well why was that?”
JENNIFER COOK
So now take me to behavioural economics. What are you trying to do? It’s building a bridge, isn’t it, between science and everyday life and theory between the social sciences?
NIKOS NIKIFORAKIS
I think perhaps sometime in the future we will cease talking about behavioural economics and we’re just going to talk about economics. One would argue that maybe we’re even going to stop talking about economics and talk about social sciences. So the idea is that there’s a constant interplay between theory and data. So we have our theory, perhaps based on observations, perhaps an introspection, we generate precise predictions and then we test them.
JENNIFER COOK
So you’re seeing it as a natural development?
NIKOS NIKIFORAKIS
Absolutely, it’s like one of my colleagues in the US, Dan Friedman, has said, “It’s the engine of scientific progress.”
So we test it and if something is not working then we adjust the theory, we have new predictions, we test them again, if they’re accurate, great, we keep testing the engine and see if it’s working. If not, we just fix it again.
JENNIFER COOK
Is this ivory tower?
NIKOS NIKIFORAKIS
I would say it’s the exact opposite. The background is to some extent, so we have our theory, it’s a very organised way of viewing the world but at least my point of view is that theory should not only be to understand individuals in an abstract world but we actually should be able to predict what’s going to happen. My goal is to have a social science that is as predictive as natural sciences.
JENNIFER COOK
All right, well that’s a wonderful segue into what you’ve been doing in the lab. Tell us about your experiments and you’ve been doing a lot of work understanding punishment and reward and how they affect behaviour.
NIKOS NIKIFORAKIS
Yes, I'm dealing with social behaviour and in particular I'm interested in social dilemmas. Now [a] social dilemma is a situation where there is a conflict between the private and the public interest. There are many situations in everyday life that we have such attention. One nice example is if you can imagine two students having to do an assignment for their class and they’re going to get the same grade. Now we both want to work hard to get a good grade, but at the same time I’d really rather you did most of the work while I have a drink with my friends.
JENNIFER COOK
Excellent.
NIKOS NIKIFORAKIS
So the question of course is how do we solve this problem? Of course economic theory would predict that since we’re all self interested, we’re just going to go and have a drink and let our friend, who has no option left, do the literature review and everything else which is needed.
So the question is how do we now enforce cooperation and over the years there have been many solutions to this problem. One that has received a lot of attention is the use of peer punishment, so that means if you go and have a drink I'm going to hit you and of course why would you do it unless you’re feeling pretty upset. So this is kind of against self interest hypothesis because it’s not nice to fight, it’s costly for you, you’re going to lose a friend and so on.
JENNIFER COOK
We’re tapping into unwritten social norms, aren’t we? All of those gradients of expected behaviour that often aren’t spoken about but are just assumed and often we get so very offended when they’re not met.
NIKOS NIKIFORAKIS
That’s absolutely right. One part is what I said, the social dilemma. Now the other thing which I talked about, punishment and enforcement, has exactly to do with what you mentioned. The problem of how to enforce cooperation arises when there’s no external authority or when there is no contract, so I cannot go to my teacher and prove that I did all the work, the other guy should get a zero.
Similarly, the problem arises when just going to the teacher, even if I could prove what I said, it’s very costly to do so because other people maybe will never work with me again. So in all these instances I'm left with a problem.
JENNIFER COOK
So it’s very fraught, there’s so many different paradigms operating in a seemingly quite simple scenario.
NIKOS NIKIFORAKIS
That’s right, and you can also imagine that there are many day-to-day interactions that fall in this category where we cannot have someone else enforce cooperation and cooperation would make us all better off. But if we’re selfish then at the end we’re both going to be worse off. Some people refer to this as a prisoner's dilemma.
Some people have been arguing that if we allow people to punish at their own expense, what we call free riders, then they will. Now to some this is a surprise because punishing is costly. And of course if you’re self interested you shouldn't do it. But of course we know that people have emotions, they’re angry and they do punish those free riders, those who go to have a drink.
JENNIFER COOK
That takes me back to an old village mentality, you put someone in the stocks and you pelt them with fruit, so I’d be really interested to see how this works.
NIKOS NIKIFORAKIS
Well that’s a very good analogy. What the lab experiments add to this knowledge is the fact that when you go to the village you can never be sure why the people throw tomatoes to the free rider. It could be that they’re angry, but it could also be that they want to make sure that this person won’t do this thing again to you in the future. So it’s strategic motivation behind this behaviour.
In the lab we can make sure that people don’t know with whom they’re interacting and they never will meet each other. So if you do take such a costly action then we know that this is something it’s just making you angry.
Some people have been arguing that if we allow people to do this, then this actually will help cooperation and one of the things I’ve found out in my research is that of course if we allow punishment, you open up the door to counter punishment. So of course one would argue, “Well, you broke the norm, you free rode, you were caught red handed. You were punished for going to have a pint.”
JENNIFER COOK
You’re listening to Up Close, I'm Jennifer Cook and I'm talking with Nikos Nikiforakis about how economic theorists can better help us understand and predict human behaviour.
So if the free rider free rode in the first place, just how amenable are they going to be to say, “Oh gee, I did the wrong thing, haven’t I been bad?”
NIKOS NIKIFORAKIS
That’s a very good point. This is one of the questions I was trying to answer. Will people who have been caught red handed having a drink... will they actually retaliate? The answer is of course they will. It’s not all of them but a good proportion of them. If they’re punished they will counter punish and this is even if we’re never going to see each other again.
So the fact that you’re telling me off is not something I enjoy and therefore for a number of reasons, it could be anything, it has to do with self esteem, who knows? It’s not clear yet. People actually do retaliate so I found that the extent of counter punishment is directly related to the extent of punishment. The more heavy your punishment is, the more likely I am to retaliate and the more severe my retaliation will be.
So if you come and politely say I really didn’t like the fact that you went for a drink, you might say yeah, you’re right. But if you come and you just say to me what kind of person are you, how dare you go for a drink when we have an assignment? I might respond well, who are you to tell me what to do?
Now, of course, I'm just putting a story to the data here because what we’re doing in the lab is actually abstract scenarios where people are given money. So free riding means, basically, I'm not contributing enough to a kind of a joint account that generates money as you are. When you see how much I’ve contributed, you can actually burn some of my money. I sacrifice some of my money, it’s like two or three dollars, but that will burn six of your dollars. Which is very pleasant for me, right?
JENNIFER COOK
For you.
NIKOS NIKIFORAKIS
Yes, how would you feel if I did that to you though?
JENNIFER COOK
I’d be coming with a bit torch to your pile of money, Nikos, but that’s just me.
NIKOS NIKIFORAKIS
That’s what most people apparently do and of course the problem with this is that once you have this you open up the gate for feuds, so long lasting argument between parties that we’ve seen in history. Now one thing about feuds is we don’t actually know why they existed. This is one thing.
So what we’ve seen, I’ve done now an experiment with my colleague in England, Dirk Engelmann, where we allow people to punish each other as long as they want until they’re bankrupt which is the lab equivalent of being dead. What we did find is, to our surprise, people are extremely smart in avoiding feuds. Now you can say they’re sitting there and they have a few seconds to think before they react, it’s not a face to face interaction, so they might have a chance to think if I do that the other person is going to come back to me. We did find behaviour which suggests that people are very smart in avoiding feuds.
JENNIFER COOK
In picking their battles.
NIKOS NIKIFORAKIS
That’s right and in fact there is some evidence from anthropology that suggests that especially in small societies if people want to sanction someone they don’t just go and kill their cow, unless they have a really good or a really bad reason, but what they do is they gossip. So in that sense we find that taking the data that people do actually get angry about this type of behaviour they are happy to retaliate but they’re very clever in how they go about retaliating.
We recently finished a study with one of my colleagues in Melbourne here, Tom Wilkening, and a colleague in the Netherlands, Charles Noussair, and we asked the following question. What does it take to see feuds in the lab? Is it the fact that it’s impossible to generate feuds when we actually have no strategic reason to do it? Is it that in the villages or in the companies where we see feuds there’s so much of our identity involved and this is the reason for feuds?
Or is it that no, we can actually somehow generate feuds in the lab and by generating feuds I mean get people in the situation where they start getting angry with each other and burning their money. Remember it’s anonymous so there’s nothing problematic, we always have ethics approval.
We did this study where we actually made very cheap punishment and in fact punishing you extra after you’ve come to me is free. You can imagine that you’ve come close and you’re slapping me and it’s not very costly to slap you back a bit. What we did find is that when we did that we did observe big feuds that lasted for a long time and people did burn each other’s money and it took them a while before some, not all, stopped punishing each other.
Now of course on the one hand this means that punishment, peer punishment, is a problematic mechanism to enforce cooperation. You would rather have a third party but what is even more interesting is the fact that we actually observed a big difference in the behaviour of the two genders in our subject pool.
We have a very big subject pool but these are just students at the University of Melbourne and what we actually found is that unlike in most experiments where women are found to be more co-operative than men, in our experiment women are less co-operative than men. Now can we explain this? We cannot be 100 per cent sure at this point why this happens, but we believe it’s the presence of punishment and the threat of the feud.
What seems to be case is that the women in our sample seem to underestimate how vindictive men will be in this scenario. For this reason they just contribute less whereas the men in our sample seem to be fully aware that by not co-operating they risk a feud.
JENNIFER COOK
Interesting.
NIKOS NIKIFORAKIS
So it could be one could tell a story but this has to be proven yet that perhaps men have evolved in these situations, that they understand how to enforce cooperation, how to avoid certain dangers and perhaps for this reason women free ride in our sample and of course this generates feuds. This is still something we’re investigating.
JENNIFER COOK
You’re listening to Up Close, I'm Jennifer Cook and I'm talking with Nikos Nikiforakis about how the field of behavioural economics can help us to understand why people do the things that they do.
These experiments in the lab are all well and good, but how do you cross that great divide between the lab and the real world and making it relevant and just how predictive can these findings be?
NIKOS NIKIFORAKIS
We need to answer this in two parts. First let me tell you that people in the lab are real people, of course, and the incentives are also real. At the University of Melbourne our students have about $40 at stake and one would think that maybe you should try with even more and in fact one of my colleagues, Lisa Cameron, did a study in Indonesia where the amount at stake at some cases was one month’s salary and that was just for I believe a half an hour experiment. I think that’s as high as stakes can be.
She found that the behaviour was pretty much invariant to this so it doesn’t matter how much perhaps is at stake, it’s more the emotions that are driving our behaviour. Maybe if people had more time to think about it maybe they would act differently but they were not pushed to behave in any rash way.
Now the other thing is about the subject pool. Typically for a number of reasons we’re using students for the experiments and the main reason actually is that students are pretty smart and they can adapt quickly in the environment and understand all the fine details.
But of course it’s important to know whether students behave in a different way than say older people or younger people. The evidence we have, and we don’t have a lot of evidence yet, the evidence does seem to suggest that behaviour is similar. There are some environments in which we know there’s absolutely no difference in behaviour. Some other environments we still don’t have enough data but we seem to find a correlation between the behaviour of these people in the lab and outside of the lab.
Still more work needs to be done, but I don’t think there is a problem so much with generalising from a lab experiment to the world outside the laboratory. it’s more a problem of realising the differences that exist outside of a laboratory. As long as we can identify the things that change out there, then we should be able to generalise. If we’re taken from one situation to another which is totally different then I wouldn’t be surprised if behaviour is different.
JENNIFER COOK
This is all so intriguing, looking at science, trying to figure out why we behave the way that we do. Now if we can pull back our focus and take a wider view for a moment, if we look at organisations such as the United Nations where not every participant is equal and taking into account what you were saying about social norms and behaviour and punishment and reward, how can you explain the dynamics of an organisation like the UN in terms of your behavioural economics?
NIKOS NIKIFORAKIS
Interaction in the UN is in essence decentralised. For example if there is one country let’s say not co-operating with the others, then we don’t have an external authority that can enforce co-operating so this is, for example another case where we asked the question how do we enforce cooperation and what are the repercussions of enforcing cooperation in this set up?
If one is to generalise from economic theory or laboratory experiments to situations like the UN, one would have to be really careful to understand how the UN functions. So it’s not a case that we can just take any experiment or any theory and apply it to the UN. We really need to understand what is the structure of interaction, how strong are the agents and how weak are certain other agents.
JENNIFER COOK
That is so interesting because it brings me to a case of where economic theorists really helped governments around the world make a load of money. We had the case of the Generation 3 mobile rights which were being auctioned by governments and those auctions were developed by economic theorists.
Now the countries that worked with those theorists made a lot of money but Nikos what happened to those who just sort of observed what those countries did and thought well we can do that without paying all that money?
NIKOS NIKIFORAKIS
That’s a good example actually. So in 2000 there were a number of auctions for the Third Generation of mobile licenses. In the UK the government worked hand in hand with a number of economic theorists and experimental economists to develop the optimal design for the auctions and to test them because, as we say, there is some kind of difference right now between theory and experiments. Not always but in some cases.
The result of this process of designing carefully this auction mechanism was that the UK I believe gathered $34 billion which was a huge amount and no one expected such a success. I think it was the next auction in the Netherlands, I think a month afterwards, where they thought they will, my understanding is, take the design that was used for the same purpose in the UK, but they didn’t take into consideration the details of the market and they just applied this design and the result was that when they expected to collect $8.5 billion, instead they collected $2.5 billion. So that auction was considered a big failure.
That was not the only example so my understanding is that in Turkey there was a similar problem. They had another auction, not identical to the ones we just mentioned, and they set the reserve price inappropriately and as a result they actually lost a lot of money.
JENNIFER COOK
Nikos, thank you so much for your time today, but before we leave I’d like you to use your own predictive powers or analytical powers and let us know what’s next for you and in the field of behavioural economics.
NIKOS NIKIFORAKIS
Well I'm going to continue doing research in understanding social norms. I'm definitely interested to conduct some field experiments across different countries to see actually how behaviour in the lab correlates with behaviour in the field when it comes to punishment.
This was an experiment I’ve been designing for a number of years and I finally got, I think, the design right now. This is what I'll be doing next and I think it addresses the point of can we generalise from the lab to the field, but we can also learn about how different cultures enforce norms and why certain people are more likely to come to punish when you’re trying to enforce a certain norm.
With regards to the field of behavioural economics, I think it’s going to continue growing and eventually we’ll be using less and less the adjective behavioural and just economics will be based on assumptions that are more close to reality. Now it should be mentioned that this comes at a cost. Having more realistic assumptions in our models comes at a cost because our theory becomes more complex in terms of the maths involved.
Since my goal personally is to have economics predict behaviour, so to make it a social science, which is actually a science, I'm reminded that science comes from the Latin word for knowledge, so we need to be evolving in approaching the truth. In that sense I think we need to continue with interplay between theory and data in order to get there.
JENNIFER COOK
That sounds, Nikos, as if your experiments are going to take us a long way down the road to achieving just that so I hope we do come back and sit in this chair and we talk about economics and there is no need to mention behavioural economics. Thank you so much for your time.
NIKOS NIKIFORAKIS
Thank you, Jen.
JENNIFER COOK
We’ve just been hearing from Dr Nikos Nikiforakis, a behavioural economist from the University of Melbourne’s Department of Economics and you’ve been listening to Up Close. Relevant links, a full transcript and more information on this episode can be found at our website at upclose.unimelb.edu.au.
You may leave a comment on any episode of Up Close by clicking on the link at the bottom of the page. Melbourne University Up Close is created by Eric van Bemmel and Kelvin Param and this episode is produced by Miles Brown and Kelvin Param. Our audio engineer is Miles Brown and our theme music is performed by Sergio Ercole. I'm Jennifer Cook, until next time thank you for joining us on Up Close. Good bye.
VOICEOVER
You’ve been listening to Melbourne University Up Close, a fortnightly podcast of research, personalities and cultural offerings of the University of Melbourne, Australia. Up Close is available on the web at upclose.unimelb.edu.au. That’s upclose.unimelb.edu.au. Copyright 2009 University of Melbourne.